The Accuracy and Reliability of Automated Oscillometric Measurement of Ankle-Brachial Index in the Diagnosis of Peripheral Artery Disease

Abstract

Comparing the accuracy and reliability of ankle-brachial index (ABI) measured by an oscillometric device with the Doppler method in peripheral arterial disease (PAD). 122 patients admitted to the department of interventional radiology with PAD were studied. ABI was measured with the 2 methods in random order. After excluding the inconclusive results, Doppler ABIs were compared with the oscillometric ABIs in each limb, the reliability was evaluated by intraclass correlation coefficient (ICC), and the accuracy and consistency were assessed by receiver operating characteristic curves and Bland-Altman method. In 122 patients (244 legs), 27 legs got inconclusive oscillometric results, 4 legs got inconclusive oscillometric and inconclusive dorsalis pedis artery Doppler results, 2 legs had oscillometric cuff pressure intolerance. Using Doppler ABI <0.9 as a diagnostic threshold, compared with the high Doppler ABI, oscillometric method had a sensitivity of 57.81%, a specificity of 95.18%, the ICC was 0.626 (95% confidence interval [CI]: 0.536-0.701), Bland-Altman method showed 11/211 (5.21%) difference points outside the 95% limits of agreement. Compared with the low Doppler ABI, oscillometric method had a sensitivity of 50.66%, a specificity of 98.31%, the ICC was 0.541(95% CI: 0.483-0.630), Bland-Altman method showed 11/211 (5.21%) difference points outside the 95% limits of agreement. Using 1.00 as the diagnostic threshold and considering error oscillometric results as abnormal ABIs, the sensitivity improved to 88.05% with high Doppler ABI and 81.42% with the low. The accuracy and reliability of oscillometric ABI in patients with PAD is unsatisfied, which makes it not suitable as an alternative method in clinic disease assessment. Using 1.0 as the ABI oscillometric threshold for PAD diagnosis can improve the diagnostic value.